Abstract We propose to determine the structural and molecular basis of membrane curvature recognition using SpoVM, a highly conserved 26-residue peptide found in Bacillus subtilis (B. subtilis), as a model. During forespore formation, SpoVM exclusively binds to the convex surface of the forespore and initiates the assembly of a protein coat. In 2009, Ramamurthi et al. discovered that SpoVM uses the membrane geometry as an ultimate cue for its final subcellular localization. However, it is unclear how the nanometer-sized SpoVM (~40 for a presumed -helix) is able to recognize the slightly curved surface of the micrometer-sized forespore. Contrary to the current belief that SpoVM assumes a long straight amphipathic -helix and shallowly associates at the membrane surface, we found that SpoVM adopts a loop-helix structure that is deeply embedded in the membrane. This proposal seeks to extend the study to model systems that are similar in curvature and lipid composition to the membrane of the B. subtilis forespore in order to elucidate the molecular mechanism of SpoVM membrane curvature recognition. We hypothesize that deep hydrophobic insertion is key for SpoVM to detect small membrane curvature. While pursuing these goals, we will exploit geometrically well-defined spherical supported lipid bilayers as a new model for the curved membrane and develop an in situ NMR approach for determining membrane protein structures. The shape of cellular membranes is a well-conserved evolutionary phenotype. Membrane shape is generated and maintained by the interplay of protein-lipid and lipid-lipid interactions. The detection and remodeling of membrane shapes are part of many essential cellular processes such as endocytosis, vesiculation and protein trafficking. Understanding the molecular mechanism for the generation, maintenance, and regulation of membrane geometry is a fundamental question in biology and will open up new therapeutic opportunities.